rudachi 1.1.0 → 1.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 72d5093c7585e3ffdf4792c81489924ab73d55988f7648b0249ff4c6669e5cb0
-  data.tar.gz: e780dc4a15c8575fc1ddee73c46923a3ea2c5d48ec175f54f6455d8b8fabb542
+  metadata.gz: b2a7d72805e33fa5e12884193bf931d6bdf7ce16d6f9243022c7fcedad8c2f23
+  data.tar.gz: '0050952b57a66b96a89ed7cebf5dc51caed87a2437cf9c4dc494b62578f90f20'
 SHA512:
-  metadata.gz: 71f5f9c92949085bad0cacf1fffe6c7dfe9f3cc0974ae2ef1ce14b4888d72ba44ce9f74feba1b6937743c0c8c861a8a5cd3a1525d250ee499277a03fdf8890f8
-  data.tar.gz: aa843e075e6cc61aa6d275bfe2fe6caf80faa6cca22c41d5ee38da0ab444e1ade89402d0f14815dcc4d16c0558347b62cfa6a79c467f011243afa92b7a22424f
+  metadata.gz: a1a1799c8bcda90ec099a44fedc81f2c2018cd1fec7fbbf1065b1dcde5282ff0d427702b3f5276679adac88ebcf27cd6370f987ebb5aa751d660b96392b91c86
+  data.tar.gz: e25716b5b9b2483bb5c3433482d0b655ddb2eee944d18d1b9fdacca409aa26a8faf20ad350d478fffcbfe0f9761e867208e81a3396749b6f189195e35e4d1f03

data/README.md

@@ -1,29 +1,41 @@
 # Rudachi
-[Sudachi](https://github.com/WorksApplications/Sudachi) wrapper Gem for JRuby.
+Ruby wrapper for [Sudachi](https://github.com/WorksApplications/Sudachi).
 
-#### For Text
+#### Text
 ```rb
 Rudachi::TextParser.parse('東京都へ行く')
 => "東京都\t名詞,固有名詞,地名,一般,*,*\t東京都\nへ\t助詞,格助詞,*,*,*,*\tへ\n行く\t動詞,非自立可能,*,*,五段-カ行,終止形-一般\t行く\nEOS\n"
 ```
 
-#### For File
+#### File
 ```rb
-File.open('sample.txt', 'w') { |f| f << '東京都へ行く' }
-Rudachi::FileParser.parse('sample.txt')
+File.open('input.txt', 'w') { |f| f << '東京都へ行く' }
+Rudachi::FileParser.parse('input.txt')
+=> "東京都\t名詞,固有名詞,地名,一般,*,*\t東京都\nへ\t助詞,格助詞,*,*,*,*\tへ\n行く\t動詞,非自立可能,*,*,五段-カ行,終止形-一般\t行く\nEOS\n"
+```
+
+#### IO
+```rb
+Rudachi::StreamParser.parse(StringIO.new('東京都へ行く'))
 => "東京都\t名詞,固有名詞,地名,一般,*,*\t東京都\nへ\t助詞,格助詞,*,*,*,*\tへ\n行く\t動詞,非自立可能,*,*,五段-カ行,終止形-一般\t行く\nEOS\n"
 ```
 
 #### With [some options](https://github.com/WorksApplications/Sudachi#options)
 ```rb
-Rudachi::TextParser.new(o: 'result.txt', m: 'A').parse('東京都へ行く')
-File.read('result.txt')
+Rudachi::TextParser.new(o: 'output.txt', m: 'A').parse('東京都へ行く')
+File.read('output.txt')
 => "東京\t名詞,固有名詞,地名,一般,*,*\t東京\n都\t名詞,普通名詞,一般,*,*,*\t都\nへ\t助詞,格助詞,*,*,*,*\tへ\n行く\t動詞,非自立可能,*,*,五段-カ行,終止形-一般\t行く\nEOS\n"
 ```
 
 ## Requirements
 
-- [JRuby](https://github.com/jruby/jruby) 9.1.3.0 or later
+#### Ruby
+- Ruby 2.3.0 or newer
+- [rjb](https://github.com/arton/rjb) 1.1.1 or newer
+- [Sudachi](https://github.com/WorksApplications/Sudachi)
+
+#### JRuby
+- [JRuby](https://github.com/jruby/jruby) 9.1.3.0 or newer
 - [Sudachi](https://github.com/WorksApplications/Sudachi)
 
 ## Installation
@@ -54,11 +66,13 @@ LEGAL  LICENSE-2.0.txt	system_core.dic
 gem 'rudachi'
 ```
 
-Then run `bundle install` .
+Then run `bundle install`.
 
 3. Initialize Rudachi
 
 ```rb
+require 'rudachi'
+
 Rudachi.configure do |config|
   config.jar_path = 'sudachi-0.5.3/sudachi-0.5.3.jar'
 end
@@ -71,8 +85,6 @@ end
 4. Did it !!
 
 ```rb
-require 'rudachi'
-
 Rudachi::TextParser.parse('こんにちは世界')
 => "こんにちは\t感動詞,一般,*,*,*,*\t今日は\n世界\t名詞,普通名詞,一般,*,*,*\t世界\nEOS\n"
 ```

data/lib/rudachi/configurable.rb

@@ -7,11 +7,10 @@ module Rudachi
     private
 
     def config_accessor(name, klass:, default:)
-      attr_def = <<~EOS
+      module_eval <<~EOS
         def self.#{name}; @@#{name}; end
         def self.#{name}=(val); @@#{name} = #{klass}.new(val); end
       EOS
-      module_eval(attr_def)
       public_send("#{name}=", default)
     end
   end

data/lib/rudachi/dependencies/jruby.rb

@@ -0,0 +1,42 @@
+require 'java'
+require 'rudachi/lazy_load'
+require Rudachi.jar_path
+
+java_import 'java.lang.System'
+java_import 'java.io.PrintStream'
+java_import 'java.io.ByteArrayInputStream'
+java_import 'java.io.ByteArrayOutputStream'
+java_import 'java.nio.charset.StandardCharsets'
+java_import 'com.worksap.nlp.sudachi.SudachiCommandLine'
+
+module Rudachi
+  module Java
+    String                = ::Java::JavaLang::String
+    System                = ::Java::JavaLang::System
+    PrintStream           = ::Java::JavaIo::PrintStream
+    ByteArrayInputStream  = ::Java::JavaIo::ByteArrayInputStream
+    ByteArrayOutputStream = ::Java::JavaIo::ByteArrayOutputStream
+    UTF_8                 = ::Java::JavaNioCharset::StandardCharsets::UTF_8
+    SudachiCommandLine    = ::Java::ComWorksapNlpSudachi::SudachiCommandLine
+  end
+
+  module StreamProcessor
+    class InvalidInclusion < StandardError; end
+
+    def self.included(base)
+      raise InvalidInclusion unless base.ancestors.include?(TextParser)
+    end
+
+    def parse(io)
+      output_stream do |output|
+        take_stdin(io.to_inputstream) do
+          take_stdout(output) do
+            Java::SudachiCommandLine.main(Option.cmds(@opts))
+          end
+        end
+      end
+    end
+
+    LazyLoad.run_load_hooks(:stream_processor)
+  end
+end

data/lib/rudachi/dependencies/ruby.rb

@@ -0,0 +1,37 @@
+require 'rjb'
+require 'rudachi/lazy_load'
+
+Rjb::load(Rudachi.jar_path.to_s)
+
+module Rudachi
+  module Java
+    String                = Rjb::import('java.lang.String')
+    System                = Rjb::import('java.lang.System')
+    PrintStream           = Rjb::import('java.io.PrintStream')
+    ByteArrayInputStream  = Rjb::import('java.io.ByteArrayInputStream')
+    ByteArrayOutputStream = Rjb::import('java.io.ByteArrayOutputStream')
+    UTF_8                 = Rjb::import('java.nio.charset.StandardCharsets').UTF_8
+    SudachiCommandLine    = Rjb::import('com.worksap.nlp.sudachi.SudachiCommandLine')
+  end
+
+  module StreamProcessor
+    class InvalidInclusion < StandardError; end
+
+    TERM = ?\n
+
+    def self.included(base)
+      raise InvalidInclusion unless base.ancestors.include?(TextParser)
+    end
+
+    def parse(io)
+      ret = []
+      while data = io.gets
+        ret << super(data).strip
+      end
+
+      ret.join(TERM)
+    end
+
+    LazyLoad.run_load_hooks(:stream_processor)
+  end
+end

data/lib/rudachi/file_parser.rb

@@ -9,30 +9,33 @@ module Rudachi
 
     def initialize(**opts)
       Rudachi.load!
-
-      @output = Java::ByteArrayOutputStream.new
-      @opts   = Option.new(opts)
+      @opts = Option.new(**opts)
     end
 
     def parse(path)
-      take_stdout do
-        Java::SudachiCommandLine.main(
-          Option.cmds(@opts).push(Java::String.new(path))
-        )
+      output_stream do |output|
+        take_stdout(output) do
+          Java::SudachiCommandLine.main(
+            Option.cmds(@opts).push(path)
+          )
+        end
       end
-      @output.toString
     end
 
     private
 
-    def take_stdout
+    def take_stdout(output)
       stdout = Java::System.out
-      stream = Java::PrintStream.new(@output)
-      Java::System.setOut(stream)
 
+      Java::System.setOut(output)
       yield
-
       Java::System.setOut(stdout)
     end
+
+    def output_stream
+      Java::ByteArrayOutputStream.new.tap do |output|
+        yield Java::PrintStream.new(output)
+      end.toString
+    end
   end
 end

data/lib/rudachi/lazy_load.rb

@@ -0,0 +1,16 @@
+module Rudachi
+  module LazyLoad
+    @@hooks = {}
+
+    class << self
+      def on_load(name, &block)
+        @@hooks[name] ||= []
+        @@hooks[name] << block
+      end
+
+      def run_load_hooks(name)
+        @@hooks[name]&.each(&:call)
+      end
+    end
+  end
+end

data/lib/rudachi/loader.rb

@@ -1,5 +1,15 @@
 module Rudachi
-  def self.load!
-    require 'rudachi/dependencies'
+  class << self
+    def load!
+      if jruby?
+        require 'rudachi/dependencies/jruby'
+      else
+        require 'rudachi/dependencies/ruby'
+      end
+    end
+
+    def jruby?
+      RUBY_PLATFORM == 'java'
+    end
   end
 end

data/lib/rudachi/stream_parser.rb

@@ -0,0 +1,10 @@
+require 'rudachi/lazy_load'
+require 'rudachi/text_parser'
+
+module Rudachi
+  class StreamParser < TextParser
+    LazyLoad.on_load(:stream_processor) do
+      include StreamProcessor
+    end
+  end
+end

data/lib/rudachi/text_parser.rb

@@ -3,25 +3,29 @@ require 'rudachi/file_parser'
 module Rudachi
   class TextParser < FileParser
     def parse(text)
-      @input = Java::String.new(text)
-      take_stdin do
-        take_stdout do
-          Java::SudachiCommandLine.main(Option.cmds(@opts))
+      output_stream do |output|
+        take_stdin(input_stream(text)) do
+          take_stdout(output) do
+            Java::SudachiCommandLine.main(Option.cmds(@opts))
+          end
         end
       end
-      @output.toString
     end
 
     private
 
-    def take_stdin
+    def take_stdin(input)
       stdin = Java::System.in
-      stream = Java::ByteArrayInputStream.new(@input.getBytes(Java::UTF_8))
-      Java::System.setIn(stream)
 
+      Java::System.setIn(input)
       yield
-
       Java::System.setIn(stdin)
     end
+
+    def input_stream(text)
+      Java::ByteArrayInputStream.new(
+        Java::String.new(text).getBytes(Java::UTF_8)
+      )
+    end
   end
 end

data/lib/rudachi/version.rb

@@ -1,3 +1,3 @@
 module Rudachi
-  VERSION = '1.1.0'
+  VERSION = '1.3.0'
 end

data/lib/rudachi.rb

@@ -2,3 +2,4 @@ require 'rudachi/config'
 require 'rudachi/option/config'
 require 'rudachi/file_parser'
 require 'rudachi/text_parser'
+require 'rudachi/stream_parser'

metadata

@@ -1,16 +1,30 @@
 --- !ruby/object:Gem::Specification
 name: rudachi
 version: !ruby/object:Gem::Version
-  version: 1.1.0
+  version: 1.3.0
 platform: ruby
 authors:
 - SongCastle
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2022-04-09 00:00:00.000000000 Z
-dependencies: []
-description: Sudachi wrapper for JRuby.
+date: 2022-04-26 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: rjb
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 1.1.1
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 1.1.1
+description: A Ruby wrapper for Sudachi.
 email: "-"
 executables: []
 extensions: []
@@ -20,12 +34,15 @@ files:
 - lib/rudachi.rb
 - lib/rudachi/config.rb
 - lib/rudachi/configurable.rb
-- lib/rudachi/dependencies.rb
+- lib/rudachi/dependencies/jruby.rb
+- lib/rudachi/dependencies/ruby.rb
 - lib/rudachi/file_parser.rb
+- lib/rudachi/lazy_load.rb
 - lib/rudachi/loader.rb
 - lib/rudachi/option/boolean_option.rb
 - lib/rudachi/option/config.rb
 - lib/rudachi/option/string_option.rb
+- lib/rudachi/stream_parser.rb
 - lib/rudachi/text_parser.rb
 - lib/rudachi/version.rb
 homepage: https://github.com/SongCastle/rudachi
@@ -50,5 +67,5 @@ requirements: []
 rubygems_version: 3.0.3
 signing_key: 
 specification_version: 4
-summary: Sudachi wrapper for JRuby
+summary: A Ruby wrapper for Sudachi
 test_files: []

data/lib/rudachi/dependencies.rb

@@ -1,19 +0,0 @@
-require 'java'
-java_import 'java.lang.System'
-java_import 'java.io.PrintStream'
-java_import 'java.io.ByteArrayInputStream'
-java_import 'java.io.ByteArrayOutputStream'
-java_import 'java.nio.charset.StandardCharsets'
-
-require Rudachi.jar_path
-java_import 'com.worksap.nlp.sudachi.SudachiCommandLine'
-
-module Java
-  String                = JavaLang::String
-  System                = JavaLang::System
-  ByteArrayInputStream  = JavaIo::ByteArrayInputStream
-  ByteArrayOutputStream = JavaIo::ByteArrayOutputStream
-  PrintStream           = JavaIo::PrintStream
-  UTF_8                 = JavaNioCharset::StandardCharsets::UTF_8
-  SudachiCommandLine    = ComWorksapNlpSudachi::SudachiCommandLine
-end